Shopify Demand Forecasting: How to Predict Inventory Needs and Stop Buying Blind

The Cost of Buying Without a Forecast

Intuition-based ordering has two failure modes, and most merchants alternate between them:

Stockouts. The most visible failure. A product goes to zero, you lose sales to competitors, customers are frustrated, and you pay premium freight on an emergency reorder. Research by IHL Group estimates that stockouts cost retailers $1 trillion annually in lost sales. For a Shopify store, a single stockout on a top SKU during peak season can represent weeks of lost margin.

Overstock. Less visible but equally damaging. Cash that could fund marketing, payroll, or product development is sitting in a warehouse as slow-moving inventory. Overstock creates storage costs, increases markdown risk, and eventually becomes dead stock — inventory that never sells at full price and must be written down or discounted into a loss. Most merchants who complain about cash flow problems are unknowingly holding 30–60 days of excess inventory on their slowest-moving SKUs.

Both failure modes share a root cause: the order quantity and timing were based on a guess rather than a calculation. Demand forecasting replaces the guess with math.

Step 1: Calculate Sales Velocity Per SKU

Sales velocity is the foundation of every inventory forecast. It answers one question: how many units of this product sell per day?

Example: A product sold 210 units in the last 30 days. Daily velocity = 210 ÷ 30 = 7 units/day.

The period you choose matters. A 30-day rolling window is the right default for most products — recent enough to reflect current demand, long enough to smooth out day-to-day noise. Using a 7-day window amplifies weekly fluctuations; using a 90-day window under-weights recent trend changes like a new ad campaign or a seasonal shift.

Calculate velocity at the variant level, not the product level. A t-shirt that sells 50 units per month at the product level might break down to 20 units of size M, 15 units of size L, 10 units of size S, and 5 units of size XL. Each variant needs its own reorder point and reorder quantity because each has its own stockout risk. Managing at the product level obscures which variants are running low.

Step 2: Translate Velocity Into a Reorder Quantity

Once you have daily velocity, the reorder quantity formula is straightforward:

Breaking this down:

• Velocity × Lead time = units you'll sell while waiting for the order to arrive
• + Safety stock = buffer for demand spikes and lead time delays (see the safety stock guide for how to calculate this)
• − Current stock = what you already have, which reduces how much you need to order

Worked example: Product has a daily velocity of 7 units, supplier lead time is 14 days, safety stock is 30 units, and current stock is 45 units.

You need to order 83 units now to avoid a stockout before the delivery arrives, accounting for your safety buffer.

The reorder point (when to trigger the order) is a related but separate calculation: Reorder point = (Velocity × Lead time) + Safety stock. Using the example above: (7 × 14) + 30 = 128 units. When stock drops below 128, it's time to place the order.

Step 3: Adjust for Seasonality

A 30-day rolling velocity works well for products with stable demand. For seasonal products — outdoor gear, holiday items, school supplies — applying last month's velocity to a forecast for next month produces the wrong answer if you're crossing a seasonal inflection point.

Two practical approaches:

Year-over-year comparison. If you have 12+ months of data, compare the same 30-day period from the prior year. If sales in June last year were 40% higher than May, weight your June forecast accordingly. This captures seasonal patterns without requiring a statistical model.

Forward-looking adjustment. If you know a demand driver is coming (Black Friday, a product launch, a PR hit), apply a multiplier to your baseline velocity. A conservative approach: if BFCM typically produces 3× your daily average for 4 days, add (velocity × 3 × 4) to your forecast for the BFCM window. Order to cover that demand before the event, not after.

The goal of seasonal adjustment is not precision — it's to ensure your orders are placed far enough ahead of demand peaks that the inventory actually arrives in time. A supplier with a 21-day lead time cannot help you during a demand spike that starts in 10 days.

How EZStock Automates the Forecast

The formulas above are straightforward, but running them manually for 50–200 SKUs across multiple suppliers is where merchants stop doing it. EZStock automates the data collection and calculation, surfacing the output — suggested reorder quantities — in the place where you actually create purchase orders.

Common Demand Forecasting Mistakes

Using product-level velocity instead of variant-level. A product that averages 40 units/month might have one variant (size M, black) at 20 units/month and another (size XL, white) at 2 units/month. Ordering based on the product average will leave you with surplus XL-white and a perpetual stockout in M-black. Always forecast at the variant level.

Ignoring lead time variability. Supplier lead times are averages, not guarantees. If your supplier usually takes 14 days but occasionally takes 21, your safety stock and reorder point need to account for the 21-day scenario. Using the average lead time with no buffer means you'll stock out on the occasions when the supplier runs late — which tends to happen during peak demand periods when they're busiest.

Setting a reorder point and never updating it. A reorder point set based on last year's velocity is wrong for this year if your business has grown or seasonal patterns have shifted. Review your reorder points quarterly at minimum, or use a tool that updates velocity automatically.

Forecasting only for fast movers. Merchants typically pay close attention to their top-selling SKUs and ignore the long tail. But slow-moving products with long lead times and high unit costs are the ones most likely to produce overstock problems. Low-velocity products need conservative forecasts and tighter safety stock to avoid tying up disproportionate cash in a product that sells 3 units per month.

Conflating a stockout with a demand forecast failure. If you stocked out because you had the right forecast but didn't place the order in time — that's a process failure, not a forecasting failure. Track both: forecast accuracy (did your predicted demand match actual demand?) and order execution (did you place orders at the right reorder point?).

Frequently Asked Questions About Shopify Demand Forecasting

What is demand forecasting for Shopify stores?

Demand forecasting for Shopify stores is predicting how much inventory you'll need and when, based on historical sales velocity. The core calculation is: daily velocity (units sold ÷ days) multiplied by your supplier's lead time, plus a safety stock buffer, minus your current stock. This tells you how many units to order right now to avoid stocking out before the delivery arrives.

How do I calculate sales velocity in Shopify?

Sales velocity = total units sold ÷ days in the period. Pull your unit sales from Shopify Analytics → Products → Sales, filter by the variant and date range you want, and divide. EZStock automates this calculation per variant using a 30-day rolling window that updates daily — no manual export required.

How far ahead should I forecast inventory for a Shopify store?

Forecast at least 1.5× your longest lead time ahead. If your slowest supplier takes 30 days, forecast 45 days ahead. For seasonal products, extend to 90 days before your peak season so you have time to place larger orders while suppliers can still fulfill them. The most common mistake is looking at today's stock level instead of projecting where it will be in 4–6 weeks at the current run rate.

Can Shopify automatically suggest reorder quantities?

Not natively. Shopify tracks stock levels and can notify you when a product goes below a threshold, but it doesn't calculate how much to order. EZStock adds this: it combines 30-day rolling velocity, per-supplier lead times, and your safety stock settings to generate a suggested reorder quantity for each product that hits its reorder point.

What is the difference between demand forecasting and reorder points?

A reorder point is a trigger — "when stock drops below X, place an order." Demand forecasting determines what X should be and how large the order should be. Without forecasting, reorder points go stale as velocity changes. With forecasting (especially dynamic velocity tracking), your reorder points stay accurate as your business grows and seasonal demand shifts.

---

For the safety stock calculation that plugs into the forecasting formula, see the Shopify safety stock guide. For how to turn a forecast into a purchase order workflow, see how to create purchase orders in Shopify.